N Is there a narrow (1685)? A.V. Anisovich,1,2 V. Burkert,3 E. Klempt,1,3 V.A. Nikonov,1,2 A.V. Sarantsev,1,2 and U. Thoma1 1Helmholtz–Institut fu¨r Strahlen– und Kernphysik, Universita¨t Bonn, 53115 Bonn, Germany 2National Research Centre “Kurchatov Institute”, Petersburg Nuclear Physics Institute, Gatchina, 188300 Russia 3Thomas Jefferson National Accelerator Facility, Newport News, Virginia 23606, USA (Dated: January 31, 2017) The helicity-dependent observable E for the reaction γd → ηn(p) with a spectator proton was recentlymeasuredbytheA2CollaborationatMAMIinMainz. Thedatawereinterpretedasfurther 7 1 evidenceforanarrowresonancewithspinandparityJP =1/2+ (P11 wave). However,afullpartial waveanalysis without anynarrowresonance leadstoanexcellent description of thedata;imposing 0 a narrow resonance with the properties suggested by the A2 Collaboration leads to a significant 2 deterioration of thefit quality: there is no need for a narrow resonance. n a J A narrow structure was observed at a mass of about s A s A s A 0.5 1/2 1 3/2 1 tot 1 8 1685MeV in the γd → ηn(p) excitation function [1– 2 7]. The structure was interpreted [8, 9] as the non- 0 strange member of the antidecuplet of pentaquarks with x] spin-parity JP = 1/2+ predicted by Diakonov, Petrov, -0.5 e and Polyakov [10]. In 2012, the observations reported -1 - l in [1, 3, 6] were introduced into the Review of Particle c Properties (RPP) under the heading of a new one-star 1600 1700 1800 1600 1700 1800 1600 1700 1800 u W, MeV W, MeV W, MeV nucleon resonance N(1685) [11] but was removed from n [ the listings in the most recent issue of RPP [12]. The interpretation of the structure as narrow resonance was FIG. 1: Legendre coefficients of the angular distributions of 2 supportedbyfurtherstudies[13–16],theresultsreported σ1/2, σ3/2 [24], and σtot for the reaction γd → ηn(p) where v in [17] were ambiguous. σtot iscalculatedas(σ1/2+σ3/2)/2. Theexperimentalresults 7 (red circles) are compared to a BnGa fit without a narrow 8 However,also coupled-channel and interference effects resonance (solid curve) or a fit imposing a narrow resonance 3 of known nucleon resonances have been discussed in the (dotted curve). 6 literature to explain the narrow structure. The Gießen 0 group interpreted the narrow dip in the γd → ηn(p) 1. excitation function as N(1650)1/2− and N(1710)1/2+ distributions (five data points per energy interval) with 0 coupled-channel effect [18], Shyam and Scholten assign third-order Legendre polynomial functions and found a 7 the dip to interferenceeffects betweenthe N(1650)1/2−, narrow dip at 1650MeV in the first order Legendre co- 1 N(1710)1/2+, and N(1720)3/2+ resonances [19]; alter- efficient. They concluded: The extracted Legendre coef- : v natively, the dip could be produced to effects from ficients of the angular distributions for σ are in good 1/2 i strangeness threshold openings [20]. agreement with recent reaction model predictions assum- X ing a narrow resonance in the P wave as the origin r The narrow dip can, however, also be explained nat- 11 a urally by interference effects in the JP = 1/2− wave of this structure. In this paper we will show that their conclusions are incompatible with the data. [17, 21–23]. In [23], the precise data reported by the A2 CollaborationatMAMI [4, 5] wereused to study the Asafirststep,werepeatedthefitwithLegendrepoly- structure. It was found that it can be explained quan- nomials. Figure 1 shows the first-order Legendre coef- titatively by interference of the two nucleon resonances ficients Aσ1/2, Aσ3/2, and Aσtot as functions of the nη 1 1 1 N(1535)1/2− and N(1650)1/2− within the JP = 1/2− invariant mass for fits to the angular distributions of partial wave. Fits which included a narrow JP = 1/2+ σ1/2, σ3/2, and σtot = (σ1/2 +σ3/2)/2. The coefficients resonance returned a zero production strength. If the Aσ1/2, Aσ3/2, and Aσtot are similar to the corresponding 0 0 0 properties of the narrow JP = 1/2+ resonance as re- total cross sections, the coefficients A and A for the 2 3 ported in [4, 5] were imposed, the fit deteriorated signif- cross sections σ and σ are shown in [24]. In the 1/2 3/2 icantly. coefficient Aσ1/2 there is indeed a narrow dip at about 1 Recently, the A2 Collaboration at MAMI reported a 1650MeV. Since the JP =1/2− partial wave dominates measurement of the helicity-dependent double polariza- the reaction, significant contributions to Aσ1/2 have to 1 tion variable E of the γd → ηn(p) reaction [24] where come from the interference between the JP =1/2− par- E =(σ −σ )/(σ +σ ), with σ being the cross tial wave and P-wave contributions. Indeed, a compari- 1/2 3/2 1/2 3/2 h section for γd → ηn(p) with neutron and photon spin son of Aσ1/2 with fit results shows that a model assum- 1 aligned (helicity h = 3/2) and or opposite (h = 1/2). ing no N(1685) (Fig. 1, solid curve) does not reproduce The data show clearlythat the structure originatesfrom the narrow dip while a model which includes a narrow theh=1/2contribution. Theauthorsfittedtheangular N(1685) (Fig. 1, dotted curve) gives qualitative agree- 2 3 35 3 s p1 /2 (-1.0,-0.6) s p1 /2 (-0.6,-0.2) s p1 /2 (-0.2,0.2) s p1 /2 (0.2,0.6) s p1 /2 (0.6,1.0) s , m b (gd→h p(n)) s , m b (gd→h p(n)) 2 30 1/2 2.5 3/2 25 2 1 20 1.5 0.05 s p3 /2 (-1.0,-0.6) s p3 /2 (-0.6,-0.2) s p3 /2 (-0.2,0.2) s p3 /2 (0.2,0.6) s p3 /2 (0.6,1.0) 15 1 10 5 0.5 0 0 0 1500 1600 1700 1800 1500 1600 1700 1800 M(g p), MeV M(g p), MeV 2 s n1 /2 (-1.0,-0.6) s n1 /2 (-0.6,-0.2) s n1 /2 (-0.2,0.2) s n1 /2 (0.2,0.6) s n1 /2 (0.6,1.0) 25 4 3702..16 5723..33 3471..09 2378..98 4309..61 s 1/2, m b (gd→h n(p)) 3.5 s 3/2, m b (gd→h n(p)) 1 20 3 15 2.5 0 0.5 s n3 /2 (-1.0,-0.6) s n3 /2 (-0.6,-0.2) s n3 /2 (-0.2,0.2) s n3 /2 (0.2,0.6) s n3 /2 (0.6,1.0) 2 10 1.5 1 0 5 0.5 1600 1800 1600 1800 1600 1800 1600 1800 1600 1800 01500 1600 1700 1800 01500 1600 1700 1800 M(g N), MeV M(g n), MeV M(g n), MeV FIG.2: Excitationfunctionsσ andσ for5binsincosθ∗ for the reaction γd → ηp(n) 1(/to2p 2 ro3w/s2) and γd → ηn(pη) γFdIG→. 3η:pT(nh)e(ttootpa)lacrnodssγsdec→tioηnns(pσ)1/(2b,otσt3o/m2)foarntdhenerweacBtnioGnas (bottom 2 rows) and newBnGa fits. The data are from [24], fits. Thesolidlinesrepresentafitwithoutanarrowresonance, statisticalandsystematicerrorsareaddedquadratically. The thedashedlinesinthebottomfigurerepresentafitinwhicha solid lines represent the BnGa fit without an additional nar- narrow resonanceis imposed with theproperties given in [4]. row resonance, the dashed lines a fit in which a narrow res- onance is imposed with the properties given in [4]. The two numbersgivetheχ2contributionofthebin,theuppernumber without, thelower numberincluding thenarrow resonance. our fit without introduction of a narrow resonance. For the differential cross sections from [4, 5] and [24], the fit returns a χ2 = 1205 for 1150 data points. Ob- MAMI ment between data and prediction. These observations viously, there is no need to introduce N(1685). When arethe basis for the conjecture in [24] thata narrowres- N(1685) was enforced in the fit with properties as given onance has been observed. There are, however, a few in [4], i.e. with M =1670MeV, width Γ=30 MeV, and arguments which disagree with this conjecture. pBr(ηn)A1n/2 = a˜ [GeV−21 10−3] = 12.3 [GeV−12 10−3], ThedipinAσ1/2 isstatisticallysignificant. Relativeto the fit returned χ2 = 1834 for the 1150 data points 1 MAMI the solid line (representing the fit with no N(1685)),the from [4, 5, 24]. The χ2’s for the new data from [24] dip in Aσ1/2 has a mean deviation−0.24±0.04 andcon- are shown in Fig. 2 for each angular bin of σ1/2 for tributes1χ2 = 15.9 for two data points. However, there γd → ηn(p), the sum is χ2 = 187.9 for the fit without is a peak in Aσ3/2 as well, at the same mass and of simi- narrow resonance and 265.8 when it is imposed. 1 larsizeandshapeasthedipinAσ1/2. Thepeakdeviates Ifthe productionstrengthisfittedfreely,itreducedto fromthesolidlineby+0.25±0.041,contributesχ2 =12.7, 1.2[GeV−21 10−3] and the total χ2 improved by 12 units and is thus of similar importance as the dip. The coeffi- to 1193. This production strength corresponds to a con- cient Aσtot follows precisely the fit with no N(1685),the tribution whichis about 100times smaller than the con- 1 data are compatible with the fit, with χ2 = 2.1 for the tribution claimed in [4, 5]. Fig. 4 shows how the χ2 in- two data points. If the dip in Aσ1/2 had a physical sig- creaseswiththestrengthofanimposednarrowN(1670). 1 nificance, it should be seen in Aσ1tot with a strength as ThenewdataonE forthereactionγd→ηp(n)–with given by the dotted line. But it is not. There is hence a spectator neutron – in [24] differed significantly from the suspicion that the dip might be a statistical fluctua- first BnGa fits which were performed before the data on tion: a small change in the observable E may lead to a double-polarization observables on γp → ηp on protons disappearance of the dip and the peak. became available [25]. To explore this discrepancy, we To test this hypothesis, we performed overall fits. In includedthenewMAMIdataforηproductionoffprotons these fits most particle properties are frozen to the val- (boundindeuterons)[24]inthefits. Figures2and3show uesderivedfromfitstopionandphoto-inducedreactions that the new data can be included in the fit without off protons. For γn reactions we use the data listed in any problems, after a slight tuning of the parameters. [23] and, in addition, the new MAMI data [24]. The lat- In Table I, we show the helicity amplitudes obtained in ter data are shown in Fig. 2 and Fig. 3, the solid line is the new fit in comparison to the fit presented in [23]. 3 TABLE I: Helicity amplitudes determined from a fit without a narrow N(1685) resonance. The T-matrix couplings are the quantities which are listed in the RPP; K-matrix couplings are given in addition. The new results are compared to those obtained in [23] which are listed in small numbers. The comparison shows the impact of thenew data from [24] and [25]. − − − − N(1535)1/2 N(1650)1/2 N(1535)1/2 N(1650)1/2 T-matrix 0.093±0.009 0.032±0.006 GeV−1/2 T-matrix -0.088±0.004 0.016±0.004 GeV−1/2 [23] 0.114±0.008 0.032±0.007 GeV−1/2 [23] -0.095±0.006 0.019±0.006 GeV−1/2 p Phase 8±4◦ 7±7◦ n Phase 5±4◦ -28±10◦ [23] 10±5◦ -2±11◦ [23] 8±5◦ 0±15◦ K-matrix 0.112±0.008 0.075±0.006 K-matrix -0.160±0.030 -0.052±0.005 [23] 0.096±0.007 0.075±0.007 [23] -0.120±0.006 -0.052±0.006 2 Summarizing, we have studied the new data on the c 1800 helicity dependence of the reaction γd → ηn(p) with a 1700 spectator proton measured by the A2 Collaboration at 1600 MAMIinMainz[24]. Wecannotconfirmtheconclusions 1500 of the authors that the dip in the first-order Legendre 1400 coefficientinanexpansionoftheangulardistributionsof 1300 σ is due to a narrow JP =1/2+ resonance. First, the 1/2 1200 dip is accompanied by a peak in the first-order Legen- 11000 2 4 √(cid:190)(cid:190)(cid:190) B r 6( h n ) | A 8 1n / 2 | , G 1 e0 V -1/2 1120-3 tdhree dcoipeffiiscdieunettoofaσ3s/t2atoifsttihcaelsflaumcetusahtaiopneisnugtgheesmtinegastuhraet- ment of E. Second, a partial wave analysis without a narrow JP = 1/2+ resonance is excellent, the inclusion FIG. 4: Increase of χ2 when a narrow resonance with mass of it with the reported properties leads to a significantly M = 1670 MeV, width Γ = 30 MeV, and JP = 1/2+ is im- worse description of the data. posedasafunctionofthesignalstrengtha˜=pBr(ηn)A1n/2. The numberof data points is 1150. Comments of Bernd Krusche and Volker Metag to an early version of this comment are kindly recognized. The workwassupportedbytheDeutscheForschungsgemeinschaft The changes in the photocouplings of N(1535)1/2− and (SFB/TR110), the U.S. Department of Energy (DE-AC05- N(1650)1/2− for protons are likely due to the inclusion 06OR23177), and the Russian Science Foundation (RSF 16- of the new data on γp→ηp [25]. 12-10267). [1] V.Kuznetsov et al., Phys.Lett. B 647, 23 (2007). [14] Ki-Seok Choi et al.,Phys. Lett. B 636 253 (2006) 253. [2] I. Jaegle et al. [CBELSA/TAPS Collaboration], Phys. [15] A. Fix, L. Tiator, and M.V. Polyakov, Eur. Phys. J. A Rev.Lett. 100, 252002 (2008). 32, 311 (2007). [3] I. Jaegle et al. [CBELSA/TAPS Collaboration], Eur. [16] M. Shrestha and D.M. Manley, Phys. Rev. C 86 045204 Phys.J. A 47, 89 (2011). (2012), idem 86 055203 (2012). [4] D. Werthmu¨ller et al. [A2 Collaboration], Phys. Rev. [17] A. V.Anisovich et al.,Eur. Phys. J. A 41, 13 (2009). Lett.111, 232001 (2013). [18] V.Shklyar,H.Lenske,andU.Mosel,Phys.Lett.B650, [5] D. Werthmu¨ller et al. [A2 Collaboration], Phys. Rev. C 172 (2007). 90, 015205 (2014). [19] R. Shyam and O. Scholten, Phys. Rev. C 78 065201 [6] V.Kuznetsov et al., Phys.Rev.C 83, 022201 (2011). (2008). [7] V.Kuznetsov et al., Phys.Rev.C 91, 042201 (2015). [20] M. D¨oring and K. Nakayama, Phys. Lett. B 683, 145 [8] M. V.Polyakov and A.Rathke,Eur. Phys.J. A 18, 691 (2010). (2003). [21] X. -H. Zhong and Q. Zhao, Phys. Rev. C 84, 045207 [9] V. Kuznetsov and M. V. Polyakov, JETP Lett. 88, 347 (2011). (2008). [22] A. V.Anisovich et al.,Eur. Phys. J. A 49, 67 (2013). [10] D.Diakonov,V.Petrov,andM.V.Polyakov,Z.Phys.A [23] A. V.Anisovich et al.,Eur. Phys. J. A 51, 72 (2015). 359, 305 (1997). [24] L.Witthaueretal.,Phys.Rev.Lett.117,132502(2016). [11] J. Beringer et al. [Particle Data Group], Phys. Rev. D [25] J. Mu¨ller et al., New data on ~γ~p → ηp with polarized 86, 010001 (2012). photonsandprotons,N∗ →Nη decaysrevisted,submit- [12] C. Patrignani et al. [Particle Data Group], Chin. Phys. ted to Phys.Lett. B. C 40, no. 10, 100001 (2016). [13] R.A.Arndtet al.,PhysRev.C 69, 035208 (2004).